Cohere
Updated May 2026
Command
Canada's enterprise-first AI
Cohere was founded by Aidan Gomez — a co-author of the original Transformer paper — and two colleagues in Toronto. Unlike other frontier labs they refuse to build a consumer chatbot, focusing entirely on enterprise deployments with strong RAG (retrieval-augmented generation), strict data handling, and private-cloud hosting. Command A and Command R+ are tuned for business workloads where accuracy on your documents matters more than trivia benchmarks.
Variants
5
Best at
best-in-class RAG and reranking tooling
Max context
256K
tokens
Pricing
Command A API
$2.50 in / $10 out
per million tokens
Command R+ API
$2.50 in / $10 out
per million tokens
Rerank
$2
per 1000 searches
Enterprise
Custom
per private-cloud deployments
Variants
Sort the table by context window or cost to find the right variant. Click any version below for a battle card with ranks, pricing notes, and official links.
| Modalities | ||||
|---|---|---|---|---|
Command A command-a-03-2025 | 256K | varies / varies | 2025 | text |
Command R+ command-r-plus | 128K | varies / varies | 2024 | text |
Command R command-r | 128K | varies / varies | 2024 | text |
Rerank 3 rerank-v3 | 4K | varies / varies | 2024 | text |
Embed v3 embed-english-v3 | 512 | varies / varies | 2024 | text |
Battle card
Context rank
#1
within Command
Capability rank
#1
modalities + reasoning
Weights
Open
self-hostable if licensed
Best fights to pick
- enterprise RAG systems
- multilingual customer support
- document QA
Rankings are Tendril directory ranks, computed from the model data shown here. Public benchmark leaderboards change often, so official docs and current benchmark pages should be checked before buying or deploying.
Learn
Lessons about this model
Structured lessons that cover Command directly or put it in context alongside its rivals.
Check yourself
Quizzes
Short, mixed-difficulty quiz sets on Command and its model family.
Other Players: MiniMax, Cohere, Amazon, 01.AI, and More
5 questions
The broader landscape of model providers.
Start quiz →Hands-on
Try these prompts
Ready-made prompts that show Command at its best. Use them in your own AI workspace, then compare the output with what you learned in Tendril.
Command R+ for RAG-style grounding
BuildersCohere's Command family is explicitly tuned for retrieval-augmented generation.
Given ONLY these retrieved passages, answer the user's question. If the answer isn't supported by the passages, say 'not supported'. Cite each claim by passage number. Passages: [1] ... [2] ... [3] ... Question: What policy changes did the company announce in Q4 2025?
Enterprise-style summarization
CreatorsCohere is positioned for enterprise — test it on a dense earnings call excerpt.
Summarize this earnings call transcript into an executive one-pager with sections: financial highlights, forward guidance, analyst concerns, and CEO tone. [Paste transcript]
Cohere Rerank on search results
BuildersCohere's Rerank model is often more valuable than their generation model — feed it a list.
Rerank these 10 search results for relevance to the query 'how to request a refund for a digital purchase within 30 days'. Return them with a 0-1 relevance score each.
Print & keep
Printable reference
One-page summaries and flowcharts — great for desks, classrooms, or study sessions.
Go deeper
Official resources
Straight from the lab — docs, API references, and the chat surfaces you can try today.
Strengths
- best-in-class RAG and reranking tooling
- true enterprise data handling (SOC 2, private cloud)
- 100+ language multilingual
Limits
- no consumer chatbot
- smaller mindshare than GPT/Claude/Gemini
- benchmark scores trail frontier